Energy efficiency is a key requirement in the design of wireless sensor networks. While most theoretical studies only account for the energy requirements of communication, the sensing process, which includes measurements and compression, can also consume comparable energy. In this paper, the problem of sensing and communicating parallel sources is studied by accounting for the cost of both communication and sensing. In the first formulation of the problem, the sensor has a separate energy budget for sensing and a rate budget for communication, while, in the second, it has a single energy budget for both tasks. Assuming that sources with larger variances have lower sensing costs, the optimal allocation of sensing energy and rate that minimizes the overall distortion is derived for the first problem. Moreover, structural results on the solution of the second problem are derived under the assumption that the sources with larger variances are transmitted on channels with lower noise. Closed-form solutions are also obtained for the case where the energy budget is sufficiently large. For an arbitrary order on the variances and costs, the optimal solution to the first problem is also obtained numerically and compared with several suboptimal strategies.